Along with HbA1c, glucose variability (GV) in type 1 diabetes (T1DM) is increasingly regarded as a primary marker of glycemic control. In our past studies we identified physiological and behavioral correlates of GV and successfully tested a biobehavioral intervention reducing simultaneously HbA1c and the risk for hypoglycemia. A signature characteristic of this research was the development of, and the reliance upon, sophisticated computer models and advanced technology, e.g. the only to date computer simulator of the human metabolism accepted by the FDA as a substitute to animal studies and the new Diabetes Assistant (DiAs) platform, which was used in the first [Worldwide] trials of ambulatory artificial pancreas. We now propose to continue our investigations by focusing on the following fundamental hypothesis: Glucose variability in T1DM is triggered by behavioral events (e.g. meals, insulin injection) that challenge the metabolic system. The timing and the magnitude of the behavioral challenges, and the ability of the metabolic mechanisms to absorb them, determine the magnitude of GV. This process develops in a certain time frame, and can be accelerated by inadequate treatment, or attenuated by precise timing and dosing of bio- behavioral control. The testing of this hypothesis will proceed in three phases: Phase 1 (triggers of GV) will use our extensive database to design an Integrated Model of Glucose Variability (IMGV), which will: (I) help understand the timing and the magnitude of system destabilization with mistimed or unbalanced treatment, and (ii) assist the design of an optimal biobehavioral intervention to be employed in Phase 3 by extensive in silicon experiments clarifying the timing and the precursors of system stabilization. Phase 2 (mechanisms of GV) will use a combination of diaries, continuous glucose monitoring, and insulin pump records in the field, augmented by hospital-based human lab studies, which will clarify key relationships between behavioral challenges, physiological glucoregulatory mechanisms, and GV. Phase 3 (control of GV) will test, using DiAs in a randomized cross-over study, the effectiveness of a novel stepwise biobehavioral intervention designed to gradually attenuate GV. This treatment will be tested in insulin pump users at poor glycemic control (HbA1c>8.0%) and/or at risk for hypoglycemia (those with history of severe hypoglycemia), and is expected to be superior to standard pump therapy in terms of reduction of GV. In summary, this project will demonstrate that: (I) in silico experiments combined with observational and human lab studies enhance hypothesis formulation/testing, and result in efficient treatment design; (ii) cutting-edge artificial pancrea technology translated for everyday use is superior to state-of-the art current treatments.